Balanced Annihilation 7.31

[Regret]

Unit/Comm napping has been and is a part of BA's gameplay since it got released. It won't go away

he didnt ask to remove it.

That's the reason I don't see why such stupidity as enemy unit napping couldn't be ruled out.

I got the impression he did.

[Beherith]

1) models (more specifically - reducing model piece number)
2) widgets (ie/ MM widget creates a large draw on the network)
3) creating new units that are scaled versions of others (ie/ supernano that is 10x better than standard nano, will reduce cpu calculation time esp. on larger games were 20% of units are nanos (ie/500+ nanos on map). Also a superfighter could be included in this to reduce lag. This however will not fix mindless spam like AK's etc.

I intend to focus on performance for the next release. 1) is already very promising, reduces the draw time of fighters, bladewings to less than half. AK's are also extremely wasteful, since they have 27 pieces, out of which only 15 are animated, a 60% increase in frame rate can be obtained just by changing these units. I am also going through often spammed units to reduce needless piececount.

2) Got some promising leads, will uncover more of this.
3) That is not planned.

[Regret]

I intend to focus on performance for the next release. 1) is already very promising, reduces the draw time of fighters, bladewings to less than half. AK's are also extremely wasteful, since they have 27 pieces, out of which only 15 are animated, a 60% increase in frame rate can be obtained just by changing these units. I am also going through often spammed units to reduce needless piececount.

2) Got some promising leads, will uncover more of this.
3) That is not planned.

Isn't Spring's bottleneck x calculations rather than graphics? As in million billion LOS circles from fighter screens.

[oksnoop2]

Well, I imagine it would take less to time calculate 15 pieces compared to 27.

[knorke]

i guess lots of pieces can mean lots of calculations (moving, rotating them etc)
though i made units with almost 100 pieces without clearly noticing slowdown compared to units with less pieces. (and my computer is 7 years old)
I think most fps killer is
-LOS
-pathfinder
-particle count
-bugs in widgets/gadgets (like doing calculation in DrawScreen instead of in GameFrame or too often, not so much a problem in BA)
-maybe collisions

[Nixa]

I intend to focus on performance for the next release. 1) is already very promising, reduces the draw time of fighters, bladewings to less than half. AK's are also extremely wasteful, since they have 27 pieces, out of which only 15 are animated, a 60% increase in frame rate can be obtained just by changing these units. I am also going through often spammed units to reduce needless piececount.

2) Got some promising leads, will uncover more of this.
3) That is not planned.

Isn't Spring's bottleneck x calculations rather than graphics? As in million billion LOS circles from fighter screens.

I believe that this is the greatest factor in the cause of lag (in larger games), and it needs to be addressed.

3) was just a indication of how it could be done, simple and effective without altering gameplay dynamics. It's a shame you don't want to address this problem

[Pxtl]

Is piececount really that big of a load? It doesn't seem like that should be the case... but I wouldn't be surprised if it is.

[Nixa]

Is piececount really that big of a load? It doesn't seem like that should be the case... but I wouldn't be surprised if it is.

Try it yourself

Simple test: give 1000 aks, zoom out, move aks, zoom in, record fps, repeat with stumpys

[Wombat]

decreasing animated parts of ak and flea gives best effects tbh. also i hope there will be no uber-sexy-epic effects for these.

also easiest way to decrease lag in later parts of the game is to replace current nano effect. reducing nano-particle-something (forgot name) makes huge difference. so making simple nano effect (even like ca one) should give nice result

fighters lag most when they all suddenly change direction, not when patroling so its not THAT bad

[Nixa]

decreasing animated parts of ak and flea gives best effects tbh.

What, when, how do you know this? If you have a fixed lag version of BA please post it!

also easiest way to decrease lag in later parts of the game is to replace current nano effect. reducing nano-particle-something (forgot name) makes huge difference. so making simple nano effect (even like ca one) should give nice result

My comment about nanos had nothing to do with particle effects, reducing those will net you almost zero gain (as particles are limited by default anyway).

fighters lag most when they all suddenly change direction, not when patroling so its not THAT bad

Umm, this is a joke right...

And if in some world you were correct with this (which you aren't), isn't changing direction when you need the most control (fps) anyway?

I do hope you are just trolling with post, because almost all of it is wrong...

We spend hours and hours going through working out the performance of certain aspects of BA. Then you come along with nothing but a random opinion and post it as a fact. Don't. It doesn't help anyone.

[Wombat]

do flea needs animated legs ? nah, we can live without them, shit is tiny anyway.

i didnt care about ur post ? i see difference between changing model and effect. and u ? and their limit is set in spring options, not 'by default'.

and no

[Nixa]

and their limit is set in spring options, not 'by default'.

Even at max they cause me no lag

[Wombat]

me

lucky us we all got same pcs.

no wait


random opinions ? these things were said long time ago rofl

[Nixa]

me

lucky us we all got same pcs.

no wait

No seriously, go try it lol. You'll lose barely any fps

[Wombat]

i tried and it increases fps. main thing i suggest when ppl complain about lags and it always helps ^^

[Nixa]

Guess I live in the 21st century then. Maybe it's time to return to the 90's with 90's technology. Is it really that hard to render 20000 green balls moving in a straight line with limited lifespan?

[Wombat]

first of all, not everone buy their pc just to play game with epic eyecandy graphics.

secondly, if my pc is good enough to run crysis i see no reason why i should get 1 fps after some noob starts to spam 500 aks.

last but not least, im not the one who started performence topic, what means im not the only one with such problems, stop being so egoistic

[Nixa]

Good enough to run crysis but not good enough to render 20000 small green balls... very interesting...

[Wombat]

my point, isnt that strange ?

[HectorMeyer]

come along with nothing but a random opinion

This is my cue to add something to the discussion too:

Isn't the real problem LOS checks, but checks for units in weapon range?

For LOS updates one could maybe reduce their update frequency without any big downsides, for example dynamically as the game progresses. For weapon range checks, maybe they can be done "collectively" under some conditions, i.e. the engine detects if there is a group of fighters on patrol, and then removes individual checks from each unit, but defines an area for which checks are done, and only does the individual checks again when an unit enters this area.

disclaimer: this is coming from someone who has no idea how the spring engine works and who's computer expertise culminates in making bold text with html